Finding, reading, cleaning and responsibly reusing the data models learn from.
Cleaning, encoding and the decisions that quietly determine how well anything trains.
Guidelines, agreement and the quality controls that make labels worth training on.
Dividing data so a score means something, and finding the leak when it does not.
Writing the card that lets somebody else use data correctly — including you, later.
Read a dataset card, load the data with pandas, deal with missing values, and check you are allowed to use it.
30 in this topic
The questions a dataset card should answer — what one row is, where the data came from, its licence and its quirks — before you use it.
Data for AI 1 min read 6 Apr 2026 For people
Accuracy, completeness, consistency, timeliness, validity and uniqueness: a framework for checking whether data is fit for purpose.
Data for AI 2 min read 5 Apr 2026 For people
Why data goes missing, how to find out, and the options — dropping, imputing, flagging — with their trade-offs.
Data for AI 2 min read 4 Apr 2026 For people
How to spot unusual values, decide whether they are errors or genuine extremes, and treat them appropriately.
Data for AI 2 min read 3 Apr 2026 For people
How to create high-quality labels: clear guidelines, measuring agreement, handling ambiguity and reducing cost.
Data for AI 2 min read 2 Apr 2026 For people
Delimiters, headers, quoting, encodings and types: how to read and write CSV files without the usual surprises.
Data for AI 2 min read 1 Apr 2026 For people
How the Parquet columnar format saves space and speeds up analysis, and when to choose it over CSV.
Data for AI 2 min read 31 Mar 2026 For people
Where to find reliable public datasets for learning and projects, and how to judge whether a dataset is suitable.
Data for AI 1 min read 30 Mar 2026 For people
CC0, CC BY, CC BY-SA, ODbL, non-commercial terms and government licences: what common data licences allow.
Data for AI 2 min read 29 Mar 2026 For people
A step-by-step checklist for getting to know a new dataset before modelling or reporting.
Data for AI 1 min read 28 Mar 2026 For people
How to clean data reproducibly: fixing types, standardising values, removing duplicates and documenting every step.
Data for AI 2 min read 27 Mar 2026 For people
Inner, left and outer joins, matching keys, and the classic mistakes — duplicated rows and silent data loss — to avoid.
Data for AI 2 min read 26 Mar 2026 For people
Why datasets need versions like code does, and practical approaches from snapshots to dedicated tools.
Data for AI 2 min read 25 Mar 2026 For people
Why training data must reflect the people and situations a model will face, and how to check and improve coverage.
Data for AI 2 min read 24 Mar 2026 For people
Creating varied training examples from existing ones — for images, text, audio and tabular data — and when it helps.
Data for AI 1 min read 23 Mar 2026 For people
What a feature store is, the problems it solves — reuse and training-serving consistency — and whether you need one.
Data for AI 2 min read 22 Mar 2026 For people
How synthetic data is generated, where it helps — testing, sharing, rare cases — and the limits on privacy and realism.
Data for AI 2 min read 21 Mar 2026 For people
How the world changes under a deployed model, how to detect drift, and what to do about it.
Data for AI 2 min read 20 Mar 2026 For people
Consent, transparency, minimisation and fair treatment of participants when collecting data for analysis and AI.
Data for AI 2 min read 19 Mar 2026 For people
A structured way to document how a dataset was created, what it contains and how it should be used.
Data for AI 2 min read 18 Mar 2026 For people